Parent pre- and post-intervention questionnaires

The researcher did not find any suitable questionnaire regarding music experiences of young children with CI; therefore, pre- and post-intervention questionnaires were designed to be completed separately by each parent (Appendix I & J). The purpose of the pre-intervention questionnaire was to provide information on the child’s music-related behaviors/experiences at home since the child’s CI implantation. The purpose of the post-intervention questionnaire was identical to the pre-, except that the parent had to take into account the time that had elapsed since the music therapy intervention and the play sessions.

Several general guidelines, recommended in the literature (Babbie, 1998; Beyth-Marom, 1986; Neuman, 2006; Robson, 2002; Sudman & Bradburn, 1983; Wigram, 2005), relate to the ways researchers may construct questionnaires. These guidelines and their incorporation into the questionnaires in this study are detailed below.

• General questionnaire format: The format, or the overall physical layout of a questionnaire, is considered to be just as important as the nature and wording of the questions since it can make a questionnaire easier to complete;

therefore, the questionnaires in this study were written in a clear, uncluttered format.

• Contingency questions: Contingency questions²⁶ were used to facilitate the parents’ task in completing the questionnaires so that the parents would not be faced with trying to answer questions irrelevant to them.

• Matrix questions: “A matrix question is a compact way to present a series of questions using the same response categories” (Neuman, 2006, p. 295). When

25 Dr. Winnie Dunn, The University of Kansas Medical Center, Kansas City, Kansas (personal communication, February 4th, 2003). E-mail: wdunn@kumc.edu

26 Contingency question: A two-part question in which a respondent’s answer to a first question directs him/her to the next questionnaire item or to a more specific and related second question (Neuman, 2006, p. 286).

Likert responses categories are used, some questionnaires employ a matrix of items. Although this format has some advantages such as using space efficiently, and enabling the respondents to complete the questionnaire faster, a disadvantage is that a matrix runs the risk of fostering a response-set among some respondents. Matrix questions were not used in the questionnaires in this study.

• Order of items in a questionnaire: The order in which questionnaire items are presented can also affect responses. The researcher was aware of this issue and made every effort to arrange the questions in a way that would be both appealing to the respondents as well as provide necessary material for the researcher.

• Introduction and instructions: Where appropriate, every questionnaire should contain clear instructions and introductory comments, such as were incorporated in these questionnaires.

• Pre-testing the questionnaire: No matter how carefully one designs a data-collection instrument such as a questionnaire, there is always the possibility of making mistakes: asking an ambiguous question, one that people cannot answer, or some other violation of the guidelines just discussed. Pre-testing the questionnaire in full and/or in part with representatives of the intended population can mitigate the effects of these pitfalls. In order to assess the design and the reliability of the pre- and post-intervention questionnaires in this study, they were administered to the family of a child with CI with whom the researcher was working at “Micha” Haifa prior to the commencement of this study as a trial subject (see first version in Appendix H). The parents answered the questions and made comments and the questionnaires were revised and refined accordingly. They were then returned to the parents for final review (see Appendices I & J for final versions).

In asking questions in a questionnaire, researchers have two options: They may use open-ended questions in which case the respondent provides his/her own answer to the question, or closed-ended questions, in which the respondent is asked to select an answer from among a list provided by the researcher. Each form has advantages

and disadvantages²⁷. A researcher’s choice to use an open or closed-ended question depends on the purpose and the practical limitations of a research project. Closed-ended questions are popular since they require little time from the respondents, provide a greater uniformity of responses and are more easily processed by being transferred directly into a computer format. Open-ended responses enable freedom and spontaneity and give opportunity to probe but they must be coded before they can be processed for computer analysis. This coding process is time consuming, often requires that the researcher interpret the meaning of responses, opening the possibility of misunderstanding and researcher bias. In addition, there is a risk that some respondents will give answers that are irrelevant to the researcher’s intent. Beyth-Marom (1986, p. 39) indicates that usually, people tend to give poor answers to open-ended questions, and many respondents do not reply to these questions. In this study, both types of questions were used, in order to encourage the parents to report about common music-related behaviors/experiences regarding their children as well as enable them to share other unique behaviors/experiences.

When constructing closed-ended questions, two structural requirements were recommended and followed in this study: (a) The response categories provided should be exhaustive by including all the possible responses that might be expected. This can be ensured by adding a category labeled “Other (Please specify: ....)”, and (b) The response categories must be mutually exclusive, which means that they do not overlap. In an attempt to follow these requirements, the researcher carefully constructed each set of response categories. In general, the response categories or choices should be balanced. Offering equal polar opposites at each end of a continuum can do this.

In addition to the general guidelines, other specific recommendations related to asking questions were described in the above mentioned literature and were applied in this study as well:

• Relevance of questions: Questions should be relevant to those who are going to respond to them. The questions in the questionnaires were directed to the parents and related to their child’s music-related behaviors/experiences at home; therefore, the questions were relevant and within the parents’

capability.

27 The advantages and disadvantages of open-ended questions and closed-ended ones are extensively detailed in Neuman, 2006, p. 286-289.

• Length and clarity of items: Long and complicated items should be avoided.

Questionnaire items were written in as clear, precise and short form as possible so that the parents did not have to cope with ambiguity, confusion and vagueness and could quickly read the item, understand its intent and easily select or provide an answer.

• Avoidance of double-barreled questions- Frequently, researchers ask respondents for a single answer to a combination of questions. As a general rule, wherever the word ‘and’ appears in a question or questionnaire statement, one should check whether he/she is asking a double-barreled question. Such questions were avoided in this questionnaire.

• Avoidance of negative items: The appearance of a negation (i.e., phrasing the question in a negative manner) in a questionnaire item paves the way for easy misinterpretation. This was avoided in this questionnaire.

• Avoidance of biased items and terms: The meaning of someone’s response to a question depends in large part on its wording. Questions that encourage respondents to answer in a particular way are biased. Words with strong emotional connotations and stands on issues linked to people with high social status can affect how respondents answer questions. Every effort was made to phrase the questions as neutrally as possible and to avoid leading questions.

• Avoidance of overlapping questions: A question that does not add information is generally unnecessary; however, some overlapping between questions is desirable and even necessary since it can serve the reliability of the measurement or its validity. In order to study the reliability of questions, sometimes one checks correlation between the answers to two questions that ask the same thing but are phrased a bit differently. When the questions are operational definitions of the same theoretical variable, it is natural that there is a partial connection between them. In such cases, the overlapping serves the aim of increasing validity. In this study, overlapping questions were not used since the questionnaires deal with facts rather than attitudes. Overlapping is more necessary in attitude questionnaires.

The questions on the questionnaires related to the following areas: the child’s responsiveness to music, the specific preferences for, or dislike of, musical sounds or

music, playing with sound games and toys, music activities or types of listening situations that enhance musical enjoyment or make music less enjoyable. Also included was a question about the parent’s expectations from this study. Question two in the pre-intervention questionnaire related to whether the family had any musical instruments at home, while the same question in the post-intervention questionnaire asked whether the family had purchased a musical instrument since this research started. Questions five and six were taken from Appendix 1 of Gfeller et al., 1999b), due to its relevance to this study. Each questionnaire consisted of 13 questions.

Embedded in each question was one or more of the following options:

• Ratings on a Likert scale²⁸;

• ‘Yes/No’ responses;

• Free-text answers.

Table 3.5 presents the various combinations of the above mentioned options for each question.

Table 3.5. Structure of parent pre- or post-intervention questionnaire

Question # Likert scale Yes/No answer Free-text answer

1 ν ν

2 ν ν ν

3 ν ν

4 ν ν

5 ν ν

6 ν ν

7 ν ν

8 ν ν

9 ν ν ν

10 ν ν ν

11 ν ν

12 ν

13 ν

28 Likert scale: “A scale, usually of approval or agreement, used in questionnaires. The respondent is asked to say whether, for example, they ‘Strongly agree’, ‘Agree’, ‘Disagree’, or ‘Strongly disagree’

with some statement” (Upton & Cook, 2002, p. 201).

Certain questions (questions #1, 2, 7, 8, 9 & 10) required the parents’ ratings on a Likert scale. Wigram (2005) indicates that “constructing categories in the Likert scale needs careful consideration” (p. 277). There is always a dilemma of how many categories to offer to the respondents. The content of a question influences the number of possible categories (Beyth-Marom, 1986). In this study, the parent had to choose one response category among five (‘Never’, ‘Seldom’, ‘Occasionally’,

‘Frequently’ and ‘Almost always’ (see Appendices I & J for definitions). Five categories, which were operationally defined in the introduction to the questionnaires, seemed sufficient to the content of the above-mentioned questions. Fink (1995) also suggests a maximum of no more than five alternative responses for self-administered questionnaires.

In addition, the researcher was aware of the fact that in questionnaires which measure attitudes, respondents tend to choose the middle response category when given an odd number of categories since it is sometimes easier not to take a position and escape to the neutral middle option²⁹. However, this study did not measure attitudes but facts, and the middle response category was not a neutral response, but rather a term that operationally defined the percentage of time the child responded in a specific manner when presented with the opportunity. Using the percentages helped focus the parent’s attention on the answer and possibly lessened the tendency to use the middle category as an easy escape.

Each questionnaire was separately completed by the mother and father in order to get each one's independent perspective. The questionnaires were completed within the two weeks prior to the beginning of the clinical trials and within the two weeks after the trials were over. Another hoped-for outcome was that the information provided could serve as a catalyst for developing a set of recommendations to parents and staff regarding the ways they could stimulate the child’s continuing growth in this area. In addition, gathering this information may have some clinical implications to share with the staff and parents as well as with other music therapy professionals.

29 It may be of interest to some researchers to track the number of respondents who choose this middle response category when given an odd number of categories. When using even-numbered sets, one advantage is that respondents are forced to commit themselves to either a positive or a negative position.